ACMA - AI-generated content

Question

1. Several jurisdictions, including the European Union, China, Spain, South Korea, Vietnam, India and the US state of California, have either mandated or are in the process of mandating the labelling of AI-generated content. a. Has ACMA considered recommending or supporting a requirement that AI-generated content be disclosed to audiences at the time it is broadcast? If not, why not? b. Does ACMA hold a view on whether Australian audiences should be able to identify, in real time, when broadcast content has been generated by AI? 2. ACMA registered an updated Commercial Radio Australia (CRA) code of practice under which the disclosure obligation for AI-generated content can be satisfied by a notice placed on a broadcaster's website or social media account, rather than at the point of broadcast. a. Does ACMA consider that a notice on a website or social media account provides effective transparency to listeners at the time AI-generated content goes to air? b. What evidence did ACMA rely on in concluding that this form of disclosure was adequate? c. Will ACMA consider imposing disclosure requirements that operate at the time of broadcast, rather than continuing to rely on industry self-regulation?

Answer

Please see the attached answer.

Documents

• Question with answer (PDF) (opens in a new tab)
• SQ26-000663.pdf (opens in a new tab)